IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308

__________________________________________________________________________________________
Volume: 03 Special Issue: 07 | May-2014, Available @ http://www.ijret.org 44
LOAD BALANCING WITH SWITCHING MECHANISM IN CLOUD
COMPUTING ENVIRONMENT

M. Aruna
1
, R. Punithagowri
2

1
Assistant Professor (Sr.G), Erode Sengunthar Engineering College, Erode, India.
2
Student, Computer Science and engineering, Erode Sengunthar Engineering College, India

Abstract
Cloud Computing provides computation, software applications, data access management and storage resources without knowing its
computing infrastructure. Cloud Computing is an emerging technology which has thousands of virtual machines consolidated to
provide different services to the end users. End users can submit jobs or tasks to the cloud environment and these jobs are termed as
Load which is executed by the cloud servers and the results are sent back to the end users. Load Balancing Algorithms are developed
to distribute the workload evenly across all the nodes so that all the nodes are available and no node should be heavily loaded.
Several Load Balancing algorithms are compared and the new approach called Switching Mechanism is applied to the partitioned
Cloud Environment.

Keywords— Cloud Computing, Load Balancing, Load Balancer, Main Controller, Partitions
-----------------------------------------------------------------------***-------------------------------------------------------------------
1. INTRODUCTION
All the large business and small business companies are
moving to cloud environment because of its scalability. The
jobs arriving to the Cloud Environment are executed by the
large data centers which have thousands of blade servers.
Cloud Computing is a Service Oriented Architecture (SOA). It
provides different types of services to the users. Users can get
the services with no need to know their infrastructure. That is,
users do not know where the service is originated and its
infrastructure. Users need to pay only for what they used from
cloud in the form of services. This is the simplicity of Cloud.

There are 4 different types of cloud environment. They are,
 Public Cloud(Free of Cost, anyone can access)
 Private Cloud(Pay for what you used, only for single
organization people)
 Hybrid Cloud(Combined both public & private Clouds)
 Community Cloud(For Communication purpose)

Users can access the cloud resources in the form of services.
There are 3 basic services provided by the Cloud
Environment. They are,
 Platform as a Service (PaaS)
 Software as a Service (SaaS)
 Infrastructure as a Service (IaaS)

1.1 Virtualization
Cloud Computing is based on the Concept of virtualization
technology. Virtualization [1] means "something that is not
real but gives all the facilities as a real one". It is the software
implementation on the bare hardware so that the resources
under the hardware can be utilized more effectively. Cloud
Computing uses the virtualization technique to make use the
cloud resources efficiently. Two types of virtualization can be
used in Cloud Environment.
1. Full Virtualization
2. Para Virtualization

1.1.1 Full Virtualization
In Full Virtualization [1], the installation of one computer is
done on the other computer. It will result in a virtual machine
that have all the facilities and softwares that are present in the
actual machine.

1.1.2 Para Virtualization
In Para Virtualization [1], the hardware allows multiple
operating systems to be run on a single machine. For this,
VirtualBox tool is used. Here all the services are not fully
available, rather than the services are provided in a partial
manner.

2. LOAD BALANCING
Load Balancing is a technique to balance the load across cloud
environment. It is the process of transferring load from heavily
loaded nodes to low loaded nodes. As a result, no node should
be heavily loaded. Thereby it will increase the availability of
nodes. If all the jobs are arrived to the single node, then its
queue size is increased and it becomes overloaded. There is a
need to balance the load across several nodes, so that every
IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308

__________________________________________________________________________________________
Volume: 03 Special Issue: 07 | May-2014, Available @ http://www.ijret.org 45
node is in running state but not in overloaded state. The goals
are as follows [2]:
 To increase the availability
 To increase the user satisfaction
 To improve the resource utilization ratio
 To minimize the waiting time of job in queue as well as
to reduce job execution time
 To improve the overall performance of Cloud
environment

Based on the current state of the system, load balancing
algorithms can be divided into two types:

2.1 Static Schemes
The current status of the node is not taken into account [3]. All
the nodes and their properties are predefined. The algorithm
works based on the predefined information. Since it does not
use current system status information, it is less complex and it
is easy to implement.

2.2 Dynamic Schemes
This type of algorithm is based on the current system
information [3]. The algorithm works according to the changes
in the state of nodes. Dynamic schemes are expensive one and
are very complex to implement but it balances the load in
effective manner.

3. RELATED WORKS
Load Balancing is one of the main issues in Cloud Computing.
Load Balancing is needed to distribute the workload to all the
servers. There are several Load Balancing algorithms have
been developed. Here, some of the Load Balancing algorithms
are compared.

Table.1 Comparison of Load Balancing Algorithms

S.No. Title Description Limitation
1. A Load
Balancing
Model based
on Cloud
Partitioning
for Public
Cloud [4]
Divide the
Public Cloud
Environment
into several
Partitions
Does not
specify the
Cloud
Division
rules and
refreshing
time period
2. Enhanced
Equally
Distributed
Load
Balancing
Algorithm
[5]
Based on the
counter
variable
allocated to
each server,
the requests
are handled
Does not
consider the
availability of
the resources
in each server
3. A New
Approach for
Load
Balancing in
Cloud
Computing
[6]
Combination
of both Min-
Min & Max-
Min
algorithms
Two
algorithms
need to be
executed
4. Power Aware
Load
Balancing for
Cloud
Computing
[7]
Executed by
Cluster
Controller,
based on
threshold
value 75%
creates new
VMs on PMs
Does not
scale up for
large data
centers
5. Efficient VM
Load
Balancing
Algorithm
for Cloud
Computing
Environment
[8]
Enhances
AMLB by
assigning
weight to
each VM &
maintains the
index table
For every
request,
parses the
index table,
thereby
increase in
response time
6. Improved
Max-Min
Algorithm in
Cloud
Computing
[9]
Assigns task
with
maximum
executing
time to
resource
produces
minimum
completion
time
Smaller tasks
has to wait
for a long
time to
execute
IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308

__________________________________________________________________________________________
Volume: 03 Special Issue: 07 | May-2014, Available @ http://www.ijret.org 46
7. Comparison
of Load
Balancing
Algorithms
in a Cloud
[10]
Cloud
Manager
receives all
requests &
places in a
queue then
check for the
availability
and allocate
to VM
Maintaining
all requests
by a Cloud
Manager
creates a
bottleneck
problem
8. Availability
and Load
Balancing in
Cloud
Computing
[2]
XMPP
servers &
XMPP clients
communicate
using MOM.
XMPP clients
sends its
current status
information to
XMPP
servers
Need to
improve
MOM for
Load
Balancing
9. Energy
Efficient
Virtual
Machine
Provision
Algorithm
for Cloud
System [11]
Proposes
Dynamic
Round Robin
Algorithm
which uses 2
rules to
consolidate
Virtual
Machines
Does not
scale up for
large data
centers

The Table.1 shows the comparison of different existing Load
Balancing techniques and their limitations. Each technique has
some limitation. Hence, the new approach need to be
developed that enhances the Cloud Environment in all aspects.

4. PROPOSED WORK
All the Load Balancing Algorithms are applied to the entire
Cloud Environment. Hence, it is difficult to manage the large
Cloud Environment. Dynamic Load Balancing algorithms
distribute the workload by checking the status of each server
which is maintained by a table called Status table. Maintaining
Status table for all the servers is a difficult one and it may not
be updated periodically hence it may not be consistent one.

The proposed work is to apply Load Balancing algorithms
only for the part of the cloud servers and not to the entire
Cloud Environment. So that, the overhead of maintaining
cloud servers is reduced. To distribute the workload in the
partitioned Cloud Environment, two main components are
needed [4]. They are,
1. Main Controller
2. Load Balancer

Before discussing about these components, the method for
partitioning the cloud environment is to be known. The entire
cloud environment is divided into partitions based on its
geographical locations. Data Centers are reside in the country
and each country has its own and unique co-ordinate value.
Based on this co-ordinate value, the Cloud Environment is
divided into partitions. The co-ordinate values are represented
as 35
O
41''22.22' N 139
O
41''30.12' E. Here, the 3 numeric
values represents the degree, minutes and seconds values and
it also represents the four directions East, West, North and
South.

The Cloud is divided into four partitions as depicted in Fig.1.
Partition-I includes the data centers that are reside in the
North-West direction. Partition II includes the datacenters that
are resided in the North-East direction. Partition III includes
the data centers that are reside in the South-West area and
Partition IV includes the data centers that reside in the South-
East area. For Example, the co-ordinate value of Tokyo is
35
O
41''22.22' N 139
O
41''30.12' E which reside in the North-
East direction, hence Tokyo belongs to the Partition-II. In the
representation of co-ordinate, it has three parts as degree,
minutes and seconds.



Fig 1 Cloud Partitions

End users from different locations submit their jobs to the
Cloud Environment. All these jobs are received by a Main
Controller which is a single node to manage all the partitions.
Nodes under each partition are managed by a Load Balancer.
Main Controller distributes the jobs to the Load Balancer by
checking its partition status. The partition may be in 3 states as
IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308

__________________________________________________________________________________________
Volume: 03 Special Issue: 07 | May-2014, Available @ http://www.ijret.org 47
Idle, Normal and Overload states. The partition status is set by
the Load Balancer based on the parameters as Number of
CPUs, the CPU processing speeds, the available memory size,
the memory utilization ratio, the CPU utilization ratio and
network bandwidth etc.,

The jobs are received by the Load Balancers and the Load
Balancing algorithms are applied to the partitions. Here the
Switching Mechanism is applied. Switching Mechanism is the
process of switching over to the 2 different algorithms
according to the 2 different situations. Switching Mechanism
contains two different algorithms, one simple algorithm for
Idle state partitions and another one effective algorithm for
Normal state partitions. Round Robin is a simple and cheap
algorithm that can be used for Idle state partitions. An
effective algorithm for Normal state partition should prevent
the partitions becoming overloaded state.

5. CONCLUSIONS AND FUTURE
ENHANCEMENT:
By partitioning method, the overhead in the management of
cloud environment is reduced. Based on the co-ordinate
values, Cloud Environment is divided into partitions. Cloud
division rule [4] issue is resolved by using co-ordinate values
for partitioning. For Idle state partitions, simple existing
algorithm can be used but for Normal state partitions, a new
algorithm is needed to be developed. In future, the concept is
to be implemented using CloudSim tool and also to propose an
effective algorithm that will work for Normal state partitions.

REFERENCES
[1] Ratan mishra & anant jaiswal, "ant colony
optimization: a solution of load balancing in cloud",
international journal OF web & semantic technology,
VOL. 3, NO. 2, PP. 33-50, 2012.
[2] Zenon chaczko, venkatesh mahadevan, shahrzad
aslanzadeh AND christopher mcdermid, "availability
and load balancing in cloud computing", international
conference ON computer AND software modeling,
VOL. 14, PP. 134-140, 2011.
[3] Venubabu kunamneni, "dynamic load balancing for the
cloud", international journal OF computer science
AND electrical engineering, VOL. 1, NO. 1, PP. 33-37,
2012.
[4] Gaochao xu, junjie pang & xiaodong fu, “a load
balancing model based on cloud partitioning for public
cloud ”, ieee transactions ON cloud computing, VOL.
18, NO. 1, PP. 34-39, 2013.
[5] Shreyas mulay & sanjay jain, “enhanced equally
distributed load balancing algorithm for cloud
computing”, international journal OF research IN
engineering AND technology, VOL. 02, NO. 06, PP.
976-980, 2013.
[6] Mohana priya.s & subramani.b, "a new approach for
load balancing in cloud computing", international
journal OF engineering and computer science, VOL. 2,
NO. 5, PP. 1636-1640, 2013.
[7] Jeffrey m. Galloway, karl l. Smith & susan s.
Vrbsky(2012), "power aware load balancing for cloud
computing", international journal OF science AND
research, VOL. 1, 2012.
[8] Jasmin james & dr. Bhupendra verma, "efficient vm
load balancing algorithm for cloud computing
environment", international journal OF computer
science AND engineering, VOL. 4, NO. 09, PP. 1658-
1663, 2012.
[9] Elzeki.o.m, reshad.m.z & elsoud.m.a, "improved max-
min algorithm in cloud computing", international
journal OF computer applications, VOL. 50, NO. 12,
PP. 22-27, 2012.
[10] Jaspreet KAUR, "comparison of load balancing
algorithms in a cloud", SINTERNATIONAL journal
OF engineering research AND applications, VOL. 2,
NO. 3, PP. 1169-1173, 2012.
[11] Ching-chi lin, pangfeng liu & jan-jan wu, "energy-
efficient virtual machine provision algorithm for cloud
system", ieee 4TH international conference ON cloud
computing, PP. 81-88, 2011.

Sign up to vote on this title
UsefulNot useful

Master Your Semester with Scribd & The New York Times

Special offer for students: Only $4.99/month.

Master Your Semester with a Special Offer from Scribd & The New York Times

Cancel anytime.